The cross-over with AI and knowledge representation is indeed one of the motivations for our amateur dabbling in cognitive science – thanks to tndal for these links.

We are of course folks looking in on this debate from the outside, trying to understand why the debate arouses such passions – a question that actually becomes more interesting, the more empirically weak the claims…

Just on an abstract conceptual level, though, there are reasons it could be hypothetically worthwhile to pursue a debate when nothing can currently be proven or disproven. The two that occur off the top of my head are (1) if the debate can begin to clarify how one might – at some future point, with better technology, for example – prove or disprove competing claims; and (2) if the debate can generate conceptual frameworks or metaphors that are useful to practitioners researching in the field, even if it seems unlikely that anyone will be able to establish the validity of the metaphors experimentally…

This doesn’t, of course, take away from your base argument that a very different line of empirical research might just obviate the whole debate. I’ll remain agnostic on this for the time being, as I can’t comment sensibly until I’ve read your materials and backtracked the discussion a bit…

Besides, I thought Chomsky’s static armchair theoretical models of human language had been supplanted by Luc Steele’s experimental models of how language develops:

http://arti.vub.ac.be/~steels/
and
http://www.csl.sony.fr/General/Publications/All.php

IMO Chomsky’s an old guy on the way out and Steels has the new experimentally proven ideas.